Engineered TnpB genome editors for plants and human cells identified by ribonucleoprotein mutational scanning

Deep mutational library construction

The TnpB reRNA DMS library was constructed from an oligonucleotide pool from Twist Bioscience and covered the 116-nt reRNA scaffold with flanking primer-binding sites for PCR amplification. The reRNA scaffold library contained ~600 variants, including all nucleotide substitutions, single-nucleotide and double-nucleotide deletions, a set of double mutations in the pseudoknot and stable tetraloop replacements in the disordered reRNA regions. The oligonucleotide library was amplified using KAPA HiFi HotStart ReadyMix with an initial denaturation at 95 °C for 3 min, followed by 16 cycles of 98 °C for 20 s, 64 °C for 15 s and 72 °C for 45 s, with a final extension at 72 °C for 1 min. The amplified library was cloned into an intermediate storage vector with NEBuilder HiFi DNA assembly master mix. The reRNA library was then assembled with a destination vector containing the variable 3′ reRNA sequence by Golden Gate cloning with BsaI-HFv2 and MlyI.

The reRNA plasmid library was digested with KpnI and ApaI and barcoded by Gibson assembly (NEBuilder HiFi DNA assembly master mix) with ssDNA oligonucleotides with internal 15 × N barcodes. The barcoded plasmid assembly was transformed into TransforMax EC100D pir-116 electrocompetent Escherichia coli and bottlenecked such that a larger culture for plasmid purification was inoculated with ~2.4 × 10⁴ transformed cells (~40 barcodes × ~600 variants), estimated from colony-forming units (CFUs) counted from titer plates. Control plasmids containing WT and catalytically dead ISDra2 TnpB were barcoded similarly.

The ISDra2 TnpB protein sequence was codon-optimized for expression in S. cerevisiae and human cells and divided into six segments of 204 bp. For each segment, mutations for all single-amino-acid changes and stop codons were designed and purchased as oligonucleotide pools from Twist Bioscience with flanking primer-binding sites, for a total of 8,116 variants. To account for enrichment of truncations at all stop codons, the nuclear localization sequence tag was positioned at the N terminus of the TnpB protein in the protein libraries, instead of the C terminus, where it was placed in the reRNA library. Mutations were designed using the most common S. cerevisiae codons, except in cases where this would create a restriction site that would interfere with library cloning or plasmid linearization. In these cases, an alternative common codon set was used to introduce the intended mutation. The first methionine was excluded from mutagenesis. The six sublibraries were amplified using KAPA HiFi HotStart ReadyMix with an initial denaturation at 95 °C for 3 min, followed by 18 cycles of 98 °C for 20 s, 64–67 °C for 15 s and 72 °C for 45 s, with a final extension at 72 °C for 1 min. Amplified sublibraries were assembled by Golden Gate cloning with BsaI-HFv2 and six corresponding intermediate cloning vectors containing the flanking WT ISDra2 sequence. Each sublibrary plasmid pool was digested with BsmBI-v2 and the concentration of the digested full-length ISDra2 protein-coding sequence for each sublibrary was measured using a Qubit 4 fluorometer. Each digested sublibrary was mixed at an equimolar ratio and inserted into the destination vector.

Single-stranded 30 × N barcode sequences were cloned into the NotI-HF-digested and XhoI-digested plasmid library with NEBuilder HiFi DNA assembly master mix. Assemblies were transformed into TOP10 electrocompetent E. coli and a larger culture for plasmid purification was inoculated with ~2 × 10⁵ transformed cells (~24 barcodes × ~8,100 variants).

Combinatorial library construction

Two combinatorial libraries with an average of ~3 and ~5 mutations per variant were created using nicking mutagenesis^25,26. DNA oligos covering 19 amino acid positions and 33 possible mutations in the TnpB protein were phosphorylated and pooled in an equimolar ratio. For synthesis of the second strand, 5 pmol and 50 pmol of the phosphorylated oligo pool was initially added with 0.38 fmol of the ssDNA template plasmid and 4.3 pmol and 43 pmol of the phosphorylated oligo pool was spiked in three times following five cycles of amplification, to generate the libraries with lower and higher mutation frequency, respectively.

RRY(N × 25)RY and YYR(N × 25)YR barcodes were cloned into the plasmid libraries with lower and higher mutation frequency, respectively, by assembly with ssDNA oligos, as described above. The barcoded library assemblies were transformed into TOP10 electrocompetent E. coli and a larger culture for plasmid purification was inoculated with ~6 × 10⁴ and ~8 × 10⁴ transformed cells for the libraries with lower and higher mutation frequency, respectively. These libraries were combined in a 1:12 ratio along with barcoded, catalytically inactivated TnpB protein controls before transformation into yeast.

Variant–barcode mapping

After library construction, variants were associated with their barcodes using long-read sequencing (PacBio Sequel II for the protein DMS library and Nanopore MinION for the reRNA DMS and combinatorial libraries). All reads were aligned to a reference plasmid and barcode sequences extracted using Minimap2 (version 2.26) and SAMtools (version 1.19)^54,55. Subalignments were made for all reads with a given barcode and a consensus sequence was created using SAMtools for all barcodes with at least two reads for PacBio sequencing and at least ten reads for nanopore sequencing. Barcodes of incorrect length and consensus sequences containing nonprogrammed mutations were discarded. For the reRNA, protein and stacked libraries, 606, 7,766 and 6,592 variants were mapped, with an average of 33, 28 and 15 barcodes per variant, respectively. Full analysis scripts and processed data are available on GitHub (https://github.com/SavageLab/tnpb_dms).

Reporter yeast strain creation

Yeast ade2⁻ reporter strains were created with the ‘delitto perfetto’ approach⁵⁶. An intermediate ADE2 KO was derived from S. cerevisiae BY4741 (American Type Culture Collection (ATCC), 201388; Meyen ex E.C. Hansen) using the CORE cassette GSKU, excluding Gal-I-SceI. To create reporter strains, the intermediate strain was cotransformed with linearized DNA containing the target site flanked by duplicate homology regions and with plasmid carrying SpyCas9 targeting the CORE cassette. SpyCas9 was constitutively expressed, triggering DSBs in the CORE cassette and repair with the linear DNA template. Target site integration was confirmed by PCR amplification and Sanger sequencing and the strain was cured of the Cas9 plasmid. Using this approach, we generated ade2⁻ reporter strains UniPAM1, UniPAM2 (target 1 strain) and UniPAM5 (target 2 strain).

Yeast pooled library selection assays

Plasmid DNA was linearized by PaqCI digestion before transformation and expression vectors were assembled by gap repair homologous recombination in yeast. Linearized plasmid libraries and backbone plasmid were transformed in a 1:3 molar ratio. For each experimental replicate, 4–5 µg of total linearized plasmid containing the DMS or combinatorial libraries of the TnpB protein was transformed. For the reRNA DMS library, 1.5 µg of total linearized plasmid library was transformed. Yeast were transformed with the lithium acetate and single-stranded carrier DNA/PEG method⁵⁷.

After transformation, cells were resuspended in synthetic complete dropout medium (SCD) lacking leucine to select for transformation and gap repair of plasmids and recovered overnight at 30 °C. The following morning, a fraction of the culture was removed for a preinduction time point. The remaining cells were induced in liquid medium lacking leucine with 2% galactose(w/v) at an initial optical density at 600 nm (OD₆₀₀) of 1.0. Induced cultures grew at 30 °C and culture samples were removed at multiple time points, pelleted, washed in milliQ water and plated on selective (−adenine −leucine) and nonselective (+adenine −leucine) SCD solid medium on bioassay dishes (Thermo Fisher). Several cell concentrations were plated at each time point to ensure maximal library coverage. Before plating, all cultures were grown with supplemental (160 mg ml⁻¹) adenine. Yeast plates were incubated at 30 °C for 48 h, after which colonies were scraped and plasmid DNA was extracted using Zymoprep yeast plasmid miniprep II (Zymo Research). Barcodes were amplified from plasmid DNA using KAPA HiFi HotStart ReadyMix (Roche) with 6–12 cycles for PCR1 and 10 cycles for PCR2. PCRs were cleaned up with Ampure XP beads (Beckman Coulter) and submitted for 150-bp paired-end sequencing on Illumina NextSeq sequencer at the Innovative Genomics Institute (IGI) NGS sequencing core.

Variant enrichment calculations

Barcode enrichment was assessed by calculating the log ratio of reads containing a given barcode in selective and nonselective samples. Barcodes with fewer than five reads in selective or nonselective conditions were removed from analysis and the log ratio was normalized by the total number of reads in selective and nonselective sequencing samples. For the protein and stacked protein libraries, variant enrichment was calculated as the median barcode enrichment for all barcodes associated with a given variant. For the reRNA library, variant enrichment was calculated as the mean of all barcode enrichments as this produced higher replicate correlation. Variant enrichments were normalized to WT such that WT had an enrichment value of zero. A two-sided Mann–Whitney test was performed to calculate the statistical significance and effect size for each variant for each replicate. Variant enrichments are represented in plots generated with DataGraph (version 5.4) and Plotly (version 5.24.1).

Yeast cleavage assays

To compare the activities of TnpB variants and orthologs, as well as of CRISPR–Cas effectors, yeast cells were transformed with 0.5–1.5 µg of clonal plasmids or linearized DNA for assembly of clonal plasmids and induced as described above. At the preinduction and postinduction time points, approximately 1–3 OD₆₀₀ units were removed from the transformed yeast culture, washed with milliQ water, resuspended in 200 µl of milliQ water and serially diluted in triplicate. Serial dilutions were plated on selective (−adenine −leucine) and nonselective (+adenine −leucine) solid SCD medium 8 h after induction, unless otherwise specified. Plates were incubated at 30 °C for 48 h, after which colony counts from serial dilutions were used to estimate the total number of CFUs. Colony reversion was calculated by dividing the number of CFUs on selective medium over the number of CFUs on nonselective medium and multiplied by 100 for the percentage. Colony reversion percentages were plotted as bar graphs with GraphPad Prism (version 10.6.1).

TnpB protein western blots in yeast

At 24 h after induction of a yeast cleavage assay, 2.5 OD₆₀₀ units of yeast cells were harvested for western blots at 3,000g for 5 min. Cells were resuspended in 100 μl of milliQ water before lysis by adding 100 μl 0.2 M NaOH and incubation at room temperature for 5 min. Cell lysate was pelleted by centrifugation at 21,000g for 2 min, washed with 200 μl of 1× PBS and pelleted again at 21,000g for 2 min. Pellets were resuspended in 30 μl of 1× PBS and 30 μl of 4× Laemmli buffer (62.5 mM Tris-HCl, pH 6.8, 10% glycerol, 1% LDS and 0.005% bromophenol blue) (BioRad). Samples were boiled at 95 °C for 3 min before 12 μl of supernatant was loaded for SDS–PAGE on a 4–20% Criterion TGX precast Midi protein gel and separated at 125 V for 60 min. Transfer to a BioRad Trans-Blot Turbo Midi PVDF transfer pack was performed with the Trans-Turbo turbo transfer system. The membrane was blocked with 5% milk in Tris-buffered saline with Tween-20 (TBST) for 1 h at room temperature and then incubated overnight with mouse anti-FLAG (1:10,000; Sigma-Aldrich, F3165, lot SLCP4941) and rabbit anti-PGK1 (phosphoglycerate kinase 1; 1:30,000; provided by J. Thorner⁵⁸) in 2.5% milk in TBST overnight at 4 °C. The membrane was washed three times with TBST for 10 min each at room temperature, then incubated with goat anti-mouse (1:10,000; LiCor, 926-32210, lot D40409-05) secondary antibody and imaged. The membrane was then washed three times and then incubated with goat anti-rabbit (1:30,000; LiCor, 926-68071, lot D40416-05) secondary antibody for 1 h at room temperature. The membrane was washed again as described before. Images were acquired on the LiCor Odyssey CLx and processed by using Image Studio version 6.1.

Comparative sequence and structure alignments

A multiple-sequence alignment (MSA) was created using an established EVcouplings (version 0.2.1) pipeline⁵⁹ with HMMER (version 3.4). The ISDra2 TnpB protein sequence was used as the Jackhmmer query for five search iterations against the UniRef90 database with a domain and sequence bit score threshold of 0.1. Redundant sequences were removed using HHfilter and sequences with less than 50% coverage were removed. Per-position amino acid frequencies were determined by calculating the ratio of the amino acid’s occurrence to the total number of sequences with a residue present at that position¹⁰.

Structural alignments of ISDra2, ISYmu1 and ISAba30 TnpBs were generated with FoldMason⁶⁰ (shown) and pairwise alignments were generated with TM-align⁶¹ with the AlphaFold2-predicted WT amino acid sequences⁶². Similarity was calculated with a Blosum62 matrix with a threshold of 1 (ref. ²⁴).

Mammalian genome editing

Mammalian cell culture experiments were performed in the HEK293T cell line (ATCC, CRL-3216; supplied by the University of California (UC) Berkeley Cell Culture Facility) and HEK293T EGFP (a gift from K. Chen, UC Berkeley)¹⁷. Mammalian cells (HEK393T or HEK293T-GFP) were grown in DMEM with high glucose, GlutaMAX supplement and pyruvate (Thermo Fisher) supplemented with 10% FBS (Avantor Seradigm) at 37 °C and 5% CO₂.

Cells were seeded at approximately 10,000 cells per well in 96-well plates 16–24 h before transfection. The transfection mix was prepared by combining plasmids encoding the protein and reRNA/sgRNA (100 ng carrying TnpB and 145 ng carrying SpyCas9 for 2.6 fmol per transfection) with 9 µl Opti-MEM I reduced-serum medium (Thermo Fisher) and 0.3 µl of TransIT-293 per transfection. Transfection mixes were incubated at room temperature for 30 min and added dropwise to the cells.

For flow cytometry, transfected plates were passaged 2 days after transfection and then harvested for flow cytometry after 2–5 days. Cells were trypsinized with 30 µl of 0.25% trypsin + EDTA was added to cells for 5 min at 37 °C and quenched with 120 µl of 1× PBS. Cells were transferred to 96-well round-bottom plates and analyzed by flow cytometry on an Attune NxT Flow Cytometer with an autosampler. Data were collected with Attune Cytometric Software (version 5.1.1) and analyzed using FlowJo software version 10.10.0 (Supplementary Data Fig. 2b).

For sequencing, cells were harvested 4 days after transfection and lysed with QuickExtract (Lucigen) according to the manufacturer’s instructions. Lysate was used directly for PCR. PCR products were cleaned with Ampure XP beads (Beckman Coulter), analyzed by a 4150 TapeStation (Agilent) and submitted for 150-bp or 300-bp paired-end sequencing on MiSeq or NextSeq sequencer at the IGI NGS sequencing core. The frequencies of the mutations were assessed by CRISPResso2 (version 2.3.1)⁶³. Editing data were plotted as bar graphs with GraphPad Prism (version 10.6.1).

Off-target analysis

To assess the specificity of TnpB and TnpB variants, CRISPR RGEN Tools (Cas-OFFinder, version 2.4.1; http://www.rgenome.net/cas-offinder/) was used to predict potential genomic off-target sites containing the ‘TTGAT’ TAM/PAM and 2–6 mismatches in the target sequence²⁷. Primers were designed using NCBI Primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/).

N. benthamiana editing

All plasmid vectors were delivered to N. benthamiana by Agrobacterium tumefaciens strain GV3101 infiltration. Cultures containing the vector of interest were grown in lysogeny broth (LB) medium supplemented with spectinomycin (50 µg ml⁻¹), gentamicin (30 µg ml⁻¹) and rifampicin (25 µg ml⁻¹) overnight at 30 °C. The next day, cultures were spun down at 3,500g for 10 min. The pellet was then resuspended in infiltration media (10 mM MgCl₂, 10 mM MES pH 5.6 and 150 µM acetosyringone, in milliQ water) and diluted to an OD₆₀₀ of 1.0. The resuspension was incubated at room temperature for 3 h before infiltration.

Syringe infiltration was performed on the abaxial surface of leaves of 4-week-old N. benthamiana plants. Infiltrated plants were then watered and transferred back to the plant growth chamber (Percival) (16-h light, 8-h dark photoperiod, 80 µmol m⁻² light intensity, 50% humidity, at 23 °C) for 4 days. After the 4-day period, leaf discs were taken for each infiltrated leaf using a hole puncher tool (Electron Microscopy Sciences, 6903950). The obtained leaf tissue was lysed in 700 µl of 2% CTAB (10 g CTAB, 100 mM Tris-HCl, 20 mM EDTA, 1.4 M NaCl, 1% polyvinylpyrrolidone) or 20 µl of Phire plant direct PCR dilution buffer (Thermo Scientific, F160S), following flash-freezing in liquid nitrogen. Leaf lysate was used directly for PCR reactions or used for genomic DNA extraction⁶⁴.

PCR reactions were performed using Phire plant direct PCR master mix (Thermo Fischer) and PCR products were cleaned with Ampure XP beads (Beckman Coulter), analyzed by a 4150 TapeStation (Agilent) and submitted for 150-bp or 300-bp paired-end sequencing on MiSeq or NextSeq sequencer at the IGI NGS sequencing core. The frequencies of the mutations were assessed by CRISPResso (version 2.3.1)⁶³.

Rice editing

Transgenic callus tissues and plants were generated by Agrobacterium-mediated transformation using established protocols⁶⁵ with minor modifications. Mature seeds of rice (O. sativa L. japonica cv. Kitaake) were dehulled and surface-sterilized for 3 min with 70% (v/v) ethanol followed by 15 min in 20% (v/v) commercial bleach (5.25% sodium hypochlorite v/v) containing one drop of Tween-20. Seeds were washed three times with sterile water to remove residual bleach. Sterilized seeds were placed on callus induction medium (CIM)⁶⁵ without BAP and incubated in the dark at 28 °C to initiate callus induction. High-quality calli were selected and transferred to fresh CIM for proliferation.

A total of 50 pieces of 6–8-week-old calli, approximately 2–3 mm in diameter, were dried on empty sterile Petri dishes for 30 min before incubation with an A. tumefaciens AGL1 suspension (OD_600nm = 0.2) carrying each transformation vector. TnpB and reRNA were cloned into the pKb-TnpB2 vector used by Karmakar et al. to deliver WT ISDra2 TnpB to rice callus³⁸. After a 30-min incubation, the Agrobacterium suspension was removed. Calli were then placed on sterile filter paper, transferred to cocultivation medium⁶⁵ and incubated in the dark at 21 °C for 3 days. Calli were then transferred to resting medium⁶⁶ (OsCIM2 supplemented with 150 mg l⁻¹ cefotaxime and 100 mg l⁻¹ timentin) and incubated in the dark at 28 °C for 7 days. Calli were then transferred to the selection medium (resting medium plus 40 mg l⁻¹ hygromycin B) and incubated in the dark at 28 °C. Tissues were transferred to the fresh selection medium every 2 weeks. The remaining callus tissues were moved to regeneration media⁶⁵ containing 40 mg l⁻¹ hygromycin B and incubated at 26 °C, under a 16-h light (90 μmol of photon per m² per s), 8-h dark photoperiod at 26 °C. When regenerated plantlets reached approximately 1 cm in height, they were transferred to 100 ml of rooting medium⁶⁵ containing 20 mg l⁻¹ hygromycin B and incubated at 26 °C under conditions of 16-h light (100–150 μmol of photons per m² per s), 8-h dark photoperiod until roots were established and leaves touched the Phytatray II lid (Sigma-Aldrich) and leaf tissue was sampled from at least four leaves from each independent transformation event for sequencing. Collected leaf samples were crushed in 50 μl of dilution buffer included in the Phire plant direct PCR kit (Thermo Scientific), then centrifuged at 18,000g in an Eppendorf centrifuge for 5 min, stored at −80 °C and used directly for genotyping.

Leaf extract was used directly for PCR with Phire plant direct PCR master mix (Thermo Scientific) according to the manufacturer’s instructions. PCR products were cleaned with Ampure XP beads (Beckman Coulter), analyzed by a 4150 TapeStation (Agilent) and submitted for 300 bp paired-end sequencing on MiSeq or NextSeq sequencer at the IGI NGS sequencing core. The frequencies of the mutations were assessed by CRISPResso2 (ref. ⁶³) and indel frequencies were averaged for leaf tissue samples originating from the same transformation event.

Pepper editing

Serrano Tampiqueno pepper (C. annuum L.) was grown in a growth chamber set at 24 °C with a 12-h light, 12-h dark cycle, a light intensity of 100 μE per m² per s and 50% humidity. The Agrobacterium GV3101 strain, carrying various TnpB and reRNA with target sequences, was grown in LB medium supplemented with SRG (spectinomycin at 50 μg ml⁻¹, rifampicin at 25 μg ml⁻¹ and gentamicin at 50 μg ml⁻¹) at 28 °C for 14–16 h. Agrobacterium cells were pelleted by centrifugation at 3,500g for 15 min, resuspended in infiltration medium containing 10 mM MgCl₂, 10 mM MES and 250 μM 3′,5′-dimethoxy-4′-hydroxyacetophenone (acetosyringone) in milliQ water to an OD₆₀₀ of 1.0 and incubated with gentle shaking for at least 3 h at room temperature. Agrobacterium was infiltrated into two fully expanded cotyledons of 20-day-old pepper seedlings using a 1-ml needleless syringe. For CaCHLHt4 and CaBRI1t11 infiltration, Agrobacterium with RNA silencing suppressor p19 was added at an OD₆₀₀ of 0.05. Infiltrated plants were kept on growth light carts and, 4 days later, leaf samples were collected and crushed in 35 μl of dilution buffer included in the Phire plant direct PCR kit (Thermo Scientific), then centrifuged at 18,000g in an Eppendorf centrifuge for 5 min and stored at −80 °C or used directly for genotyping.

Next, 2 µl of leaf extract was used for PCR with Phire plant direct PCR master mix (Thermo Scientific) according to the manufacturer’s instructions. PCR products were cleaned with Ampure XP beads (Beckman Coulter), analyzed by a 4150 TapeStation (Agilent) and submitted for 300 bp paired-end sequencing on MiSeq or NextSeq sequencer at the IGI NGS sequencing core. The frequencies of the mutations were assessed by CRISPResso2 (ref. ⁶³).

Nucleic acid and plasmid preparation

All DNA oligonucleotides used in this study, unless otherwise indicated, were synthesized by Integrated DNA Technologies. Plasmids, unless otherwise indicated, were assembled by Golden Gate cloning.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.